Birgit Betina Ertl-Wagner MD, Presenter: Nothing to Disclose

Jeffrey H. Blume PhD, Abstract Co-Author: Nothing to Disclose

Donald Peck PhD, Abstract Co-Author: Nothing to Disclose

Jayaram K. Udupa PhD, Abstract Co-Author: Consultant, Nevro Imaging, Inc

Benjamin A. Herman MS, Abstract Co-Author: Nothing to Disclose

Anthony Levering, Abstract Co-Author: Nothing to Disclose

Ilona Maria Schmalfuss MD, Abstract Co-Author: Nothing to Disclose

et al, Abstract Co-Author: Nothing to Disclose

Computer assisted systems for estimating of brain tumor volumes must function largely operator-independent and in an adequate time frame. We aimed to investigate the dependence of two software systems upon the operator’s level of professional training and the time and ease-of-use.

MR images of 25 patients with histology-proven malignant glioma acquired at two time points were analysed by 16 independent readers grouped by levels of professional training (8 MR-certified technologists, 4 neuroradiology fellows, 4 staff neuroradiologists). The software systems Eigentool and 3DVIEWNIX-TV were used to estimate the tumor volume in all cases at both time-points. Tumors were balanced for magnitude of volume change and complexity by an independent consensus panel. Estimated tumor volumes, times to train and times to estimate the tumor volumes were recorded for each case and reader. Reader satisfaction and preferences were recorded. ACRIN receives research funding from the National Cancer Institute through the grants U01 CA079778 and U01 CA080098.

The mean time to train was approximately 2 hours for either software systems. The mean time to estimate volumes of enhancing tumor and tumor plus edema was approximately 15 minutes per case for each system. Neither the level of professional expertise nor the level of complexity of the case significantly influenced the volume change estimated by the readers with either of the software systems beyond the margin of error of the methods.

Volume estimates with computer assisted software systems are largely independent of the level of the operator’s professional training making decentralized volume estimations feasible. Time to train and time to perform the volume estimation are within reasonable limits for clinical and research applications for both software platforms.

The operator independence of volume estimations of malignant gliomas and the acceptable performance times make decentralized readings feasible in the clinical and in the multi-centric research setting

Ertl-Wagner, B, Blume, J, Peck, D, Udupa, J, Herman, B, Levering, A, Schmalfuss, I, et al, , Operator Dependence and Performance Evaluation of Computer-assisted Tumor Volume Estimation from MR Images in Patients with Malignant Glioma—Results from the American College of Radiology Imaging Network (ACRIN) 6662 Trial.  Radiological Society of North America 2006 Scientific Assembly and Annual Meeting, November 26 - December 1, 2006 ,Chicago IL. http://archive.rsna.org/2006/4437665.html